Microarray analysis is perhaps the only currently available technology that can rapidly generate simultaneous expression profiles of thousands of genes. Raw microarray data represents a new class of information made available to biologists. This type of data has been exploited to measure gene response in a wide range of systems including hormone stimulation, serum response, and cancer with substantial success in measuring coordinated expression of genes across heterogeneous sample sets, a methodology known as profiling. Together, the projects that will use the Microarray Core will study skin development using mice with genetic modifications along the skin development pathway. Each modified mouse interacts with a different developmental time point and the series of mice connect early to late skin development. The set of developmental time profiles for these mice should be related at early time-points and diverge at later points where the modified genes play important roles. The impact of each modified gene will be evident in the change in downstream gene profile. Key to the understanding of the resulting profiles is appropriate treatment of this fundamentally multivariate data. The data must be screened for data quality and consistency which is enabled through replicate printing of gene targets within each array and replication of experiments across multiple arrays. Understanding of variations present in the assay is important in deciding the number of replicate experiments required for each sample. Three sources of variation can be identified and modeled: 1) biological variation (sample), 2) labeling variation, 3) chip variation (hybridization/target uniformity). The size of variation due to these sources will be estimated by the Core for the tissues under study.. From previous experience in the Core, consistency of methodology will be critical to obtain meaningful results and keep this variation low. Therefore the Core will perform all analysis steps after RNA isolation, generating labeled probe for the sample and control (pool RNA generated in the Core Facility), hybridizing and analyzing arrays. Furthermore, all samples will be compared to a standard reference pool such that the profiles can be compared to each other. Profiles will be grouped and reduced by multivariate analysis.